Several Key Technologies Of Dollar Recognition Based On Multispectral Image

In recent years,multispectral image technology has attracted more and more attention.Because of the ability to form different images under different spectra,it is widely used in the fields of bill,certificate,paper currency recognition and counterfeiting.The recognition of dollar based on multispectral image has become a hot topic in the field of paper currency recognition.Among them,the recognition of dollar denomination information,crown numbers and authenticity are very important modules in multispectral image processing.In this paper,the shortcomings of traditional recognition algorithms are found in the study and practice of US dollar based on multispectral images.This paper makes an in-depth study of US dollar towards version recognition,crowns extraction in complex background and true and false recognition of global texture features.The main work of this article includes:First,the characteristics of dollar paper money are studied,and US dollar orientation and version recognition based on Haar-Like H feature face detection are proposed,which solves the problem of instability of orientation and version recognition due to the fouling of paper money boundary,and the speed of calculation is faster.Secondly,the characteristics of the color version of the dollar crown are studied,and a crown number extraction algorithm based on the Haar-Like H feature and the extreme stable region is proposed,which solves the problem of poor extraction quality and low recognition accuracy of the crown font size under complex background texture.And other issues,but also improve the quality of the extracted font quality background.Thirdly,the anti-counterfeiting features of the U.S.dollar spectrum are studied,and the identification algorithm based on the global GLCM texture features is proposed,which solves the problems of low accuracy based on local feature pseudo-pseudo-acceptance and high computation complexity based on global features.The algorithms proposed in this paper have been successfully applied to cash registers,counters and counterfeit counters of many manufacturers.The experimental results show that compared with the traditional US dollar recognition algorithm based on multispectral image,the proposed algorithm has a higher recognition rate for the recognition and version recognition and more robust to the abnormal currency.The proposed algorithm can effectively improve the accuracy of the Wacom character recognition of the quality of the crown character extraction in the complex background.In this paper,a dollar authenticity recognition algorithm based on global GLCM texture features is proposed to improve the recognition ability of counterfeit currency and reduce false positives of real currency.